Michelle ThompsonConservationists believe an artificial intelligence (AI) tool trained to differentiate between grey and red squirrels could be a “game changer.” The system, known as Squirrel Agent, has been trained on thousands of images, enabling it to identify the squirrels with 97% accuracy.
The AI is used to control access to squirrel feeders, allowing only red squirrels to access food-containing feeders and grey squirrels to access those with contraceptive paste.
Emma McClenaghan, co-founder of Genysys Engine, which developed the tool, said, “It’s a real showcase of what AI can do,” highlighting its ability to perform a task that human volunteers cannot do at scale.
Squirrel Agent is currently being tested across the UK at several wildlife sites in partnership with five wildlife charities. Genysys Engine aims for broader use of this technology, not just for squirrels but other species that could benefit from advanced digital monitoring.
Ian Glendinning from Northern Red Squirrels, one of the conservation groups involved in the trial, emphasized the urgent need for technological help: “We are in the bar of the last chance saloon, and the landlord has just called last orders.”
Red peril
Grey squirrels are largely responsible for the significant decline in red squirrel populations. This is due to their large numbers and the fact that they carry a virus, to which they are immune but which is deadly to red squirrels. Grey squirrels arrived in the UK around 200 years ago.
“Red squirrels are holding on in Scotland and Ireland, with some island populations like Anglesey and the Isle of Wight,” explained Mr. Glendinning. “However, mainland England and Wales are in critical condition, and red squirrels will become extinct if we can’t reverse this trend.”
The technology is currently being trialed in several woodlands across the UK.
Efforts to protect red squirrels are made more difficult by the fact that not all red squirrels are actually red. Other distinguishing features, such as their tails, ears, size, and weight, can also help differentiate them from the grey squirrels, which make up the dominant population, even if their colors are similar.
The detection process is labor-intensive for human observers. However, Squirrel Agent utilizes AI to quickly analyze these distinguishing markers, accurately identifying the squirrels with a high degree of precision. Once identified, the system can automatically determine whether to help boost their survival by providing extra food or reduce their reproduction chances by directing them to contraceptives.
By a whisker
Genysys Engine plans to advance the technology by moving beyond identifying squirrel species to recognizing individual animals. Emma McClenaghan explains that this will be achieved by focusing on each squirrel’s whiskers, which are unique to every animal, much like human fingerprints.
By analyzing the tips, bases, and lengths of whiskers, the goal is for AI to help wildlife experts track and monitor specific squirrels. As the data accumulates, scientists and researchers will be able to trace family lines.
AI’s ability to rapidly process large volumes of data has already proven valuable in conservation efforts. For instance, the World Wildlife Fund (WWF) used AI to analyze millions of images of bushfire-affected areas in Australia to identify locations where animals had survived.
AI has also been embraced by hobbyists, such as a British man who created the “Furbinator 3000” to prevent foxes and badgers from soiling his garden in Surrey.
